This invention relates to reconfigurable matrix acoustical arrays, and in particular to integrated matrix arrays configured for consolidating signal paths and orthogonally reconfigurable for operating orientation.
Diagnostic ultrasound is an established and growing medical imaging modality. Currently one-dimensional ultrasound transducer arrays with up to 128 transducers are the standard in the industry. Separate coaxial cables are used to connect the transducers to the system electronics. Improved image quality requires the use of matrix (n by m) arrays with a thousand or more transducers. As transducer numbers increase and their dimensions grow smaller, limitations to present fabrication technologies arise. Cost, ergonomics, produce-ability and reliability are important issues. Signal loss due to the capacitance of the coax cables becomes a fundamental problem.
Medical ultrasound systems transmit a short pulse of ultrasound and receive echoes from structures within the body. The handheld probes are most often applied to the skin using a coupling gel. Specialty probes are available for endocavity, endoluminal and intraoperative scanning.
Almost all systems on the market today produce real-time, grayscale, B-scan images. Many systems include colorflow imaging.
Real-time images move as the operator moves the probe (or scanhead). Moving structures, such as the heart or a fetus, are shown on the video monitor.
Grayscale images depict the strength of echo signals from the body as shades of gray. Stronger signals generally are shown as bright white. Lower signals become gray and echo-free regions are black.
B-scans are cross-sectional or slice images.
Colorflow imaging adds a color overlay to the black and white image to depict blood flow.
Over the last 30 years, the major technical developments that have improved imaging or added diagnostic capability include:
Digital technology (late 70""s to early 80""s) provided image stability and improved signal processing;
Real-time imaging (late 70""s to early 80""s) provided quicker, easier imaging and functional information;
Electronically scanned linear arrays (late 70""s to early 80""s), including sequenced arrays and phased arrays, provided improved reliability;
Color-flow imaging (late-80""s) opened up new cardiac and vascular applications;
Digital beamformers (early 90""s) improved image quality;
Harmonic Imaging (late 90""s) provided improved image quality particularly in difficult to image patients;
Coded-excitation Imaging (late 90""s to present) permitted increased penetration allowing use of higher frequency ultrasound thereby improving image contrast;
Contrast agents (late 90""s to present) offer improved functional information and better image quality.
3D (volumetric) imaging (late 90""s to present) presents more easily interpreted images of surfaces such as the fetal face.
Referring to FIG. 1, there is illustrated a conventional linear transducer array ultrasound imaging system with probe shown in partial cross section, consisting of a system console [1], housing system electronics [2], to which can be connected the transducer probe assembly [4]. The probe assembly consists of the molded case [5] within which is housed an acoustic lens [6] over a 1 by 128 piezocomposite transducer array [8] with acoustical matching layers, an absorptive backing and structural support, and flexible printed circuit [10], connected to a 135 wire cable and mating connector for attaching to the system console. More details are provided below.
To form a typical sector (or wedge shaped) image, separate pulses are transmitted from each of the transducers of the array. The pulses are time-delayed with respect to each other so that the summation of the individual pulses is a maximum in the desired radial direction.
Upon reception, the echo signals received from structures within the body at each transducer are delayed with respect to each other to achieve a similar maximization along the same radial line. These signals are stored digitally.
To generate the next radial line in the image, the transmitter and receiver time delays are adjusted to change the direction of the maxima and the process is repeated. Images are thus built up line by line. Using digital storage (scan-conversion), they are converted to a conventional raster-scanned, gray-scale video image.
In general it is not required that the lines be contiguous, i.e. the selected line may come from one portion of the image on one pulse and a completely different portion of the image on the next pulse. The only requirement is that the image space is completely covered during the video image frame time. For example in a colorflow image overlayed on a grayscale image, the number of pulses allocated to the color portion may be several times that of the grayscale image.
When a pulse is transmitted by an array, transmitter time delays on each channel may also provide a focusing effect in addition to beam steering. On reception, the time delays may be adjusted in real-time as the pulse propagates into the body. This, provides a focusing effect that tracks the pulse. The dynamic, or tracking focus, thus sweeps out from the probe at the velocity of sound. Almost all ultrasound systems use dynamic focusing which provides greatly improved resolution and image quality in the scanning plane.
Referring to FIG. 2, one-dimensional (1 D, linear or 1xc3x97m) electronically scanned arrays are in widespread use today. Matrix arrays consisting of (nxc3x97m) transducers will be required in future systems to improve image quality. The various types of matrix arrays are the main topics of this discussion.
Referring to FIG. 2A, 1 D arrays may have as many as 128 transducers and either be flat or curved. All such arrays on the market today are connected to the system electronics through a bundle of coaxial cables. Beamformers in the system electronics adjust the time delays between channels to provide electronic sector scanning and focusing. High performance systems typically use all 128 transducers in their beamformers. Lower performance systems may use as few as 16 of the 128 transducers at any instant. The scanning function is performed by switching transducers into the aperture on the leading edge of the scan and switching out transducers at the trailing edge. Use of a curved array as discussed in Erikson, K. R, xe2x80x9cCurved Array of Sequenced Ultrasound Transducersxe2x80x9d, U.S. Pat. No. 4,281,550, issued Aug. 4, 1981, produces a sector scan in these simpler, lower cost arrays.
Although one-dimensional arrays are almost universally accepted, these simple linear arrays have a basic limitation on image quality due to their fixed focus in the out-of-plane or elevation dimension. This leads to a slice thickness artifact. While the images appear to be infinitely thin slices, in fact they have finite thickness that changes along the depth dimension. This poor resolution can lead to many different artifacts. The most common is the filling-in of regions where echo levels are very low, with information from surrounding tissue.
Referring to FIG. 2B, 1.25 D arrays typically use a (128xc3x973) or (128xc3x975) matrix. They are connected to the system electronics through a similar bundle of coax cables as the 1D array. The same beamformers are also used for scanning and dynamic focusing. As the pulse propagates into the body, only the center transducer is initially selected for receiving the reflected signals. By switching in additional transducers as the pulse propagates, the receiving aperture is enlarged and the receiver is weakly focused. Moderate improvements in image quality are obtained.
Referring to FIG. 2C, 1.5 D arrays use a (128xc3x97n) matrix, with n typically an odd number, typically 5, 7 or 9. 1.5 D arrays use dynamic focusing in the plane perpendicular to the scanning plane. This produces optimal resolution in all dimensions, further reducing artifacts. The key difference between the 1.25 D and 1.5 D arrays is the active time-delay beamforming in both dimensions. The number of transducers in the elevation direction is often an odd number because transducers on each side of the beam axis are electrically connected together since they both have the same time delay for on-axis targets.
Referring to FIG. 2D, 1.75 D arrays are very similar to 1.5 D arrays with the exception that the transducers in the elevation direction are individually connected to the beamformer. Limited angular beamsteering can be performed in addition to dynamic focusing. Aberration correction is also possible with the 1.75 D array. These added capabilities are not present in a 1.5 D arrays, which only provides improved focusing for on-axis targets.
Referring to FIG. 2E, 2 D arrays are the most general type, with (nxc3x97m) transducers. Dynamic focusing as well as sector beamsteering in any arbitrary direction around the axis normal to plane of the array is possible. The angles are only limited by the constraints of the beam former, the number of transducers and their dimensions.
The improvements to ultrasound arrays discussed above were examples of technology push, i.e. new technology was developed and new applications followed. Improvements in transducer and array technology were either required or enabled many of the innovations.
One of the next major innovations is expected to be the use of matrix arrays, as opposed to the linear arrays currently in use. Although the 1.25 D array produces moderate improvement, the improved image quality of 1.5 D and 1.75 D arrays will make a quantum jump in resolution, image quality and freedom from artifacts.
Referring again to FIG. 1, fabrication techniques for 1 D probes (scanheads) are well established. FIG. 1 shows a typical probe in partial cross-section. The connector [3] that couples to the system electronics [2] of console [1] is a Cannon(copyright) DL 260 pin. The cable [5] has 135 coaxial cables bundled together in a sheath. Cable weight and flexibility are important ergonomic concerns for the operators who use the scanheads daily and for extended periods. Specifications of typical high performance cables are listed in Table 1. The scanhead end of the cable is terminated in a high density, fine pitch edge connector [7], (not shown).
The multi-coax cable is electrically connected to the active piezocomposite array with a flexible printed circuit. One end of this flex circuit plugs into connector [7] and the other end is soldered or bonded with conductive epoxy to the array transducers themselves.
There are three major problems related to the use of passive cables with matrix arrays:
Although cable technology has improved dramatically in recent years, cables with thousands of coaxes are not available. Cost, weight and flexibility are issues.
Interconnecting coaxial cables to the array transducers becomes increasingly difficult.
With higher ultra sound frequencies and more complex matrix arrays, transducer size decreases. The capacitance of the transducer decreases linearly with the area of the transducer. Using a conventional cable results in a critical and fundamental problemxe2x80x94signal loss.
All linear arrays currently on the market use piezoelectric materials as the transducing mechanism from electrical signals to ultrasound (transmitter) and ultrasound back to electrical signals (receiver). The signals are generally in the form of short pulses or tone bursts.
Referring to FIG. 3, there is illustrated a graph of one way signal loss characteristics using a typical coaxial cable as a function of frequency when connected to single elements of types of arrays. The cable used for these calculations has a capacitance of 106 pF that is typical of a two meter long cable used in ultrasound systems. Signal loss occurs because of the mismatch in impedance between the array element and the cable. This signal loss applies to both the transmitting and receiving directions.
In a beam steered array, the transducer dimensions must be about a wavelength in the steering dimension. For example, in a 3.5 MHz (1xc3x97128) array the transducer width is about 0.2 mm for a total array length of approximately 64 mm. In the other dimension, the transducer dimensions are a tradeoff between resolution and depth of focus. For a 3.5 MHz array, this dimension is 12 to 15 mm.
As the frequency of the array increases, transducer size decreases, as does transducer thickness, however, the aspect ratio remains constant. Other methods of fabrication such as laser milling or scribing, etching or deposition are under development. At present, they are not well accepted.
Four curves [82], [84], [86] and [88] represent the signal loss associated with probes with no active electronics. (The additional curves, [92], [94], and [96], relate to the invention, and will be discussed in later sections.) In each case, the piezocomposite array elements have the following properties:
Relative dielectric constant, ∈=700
Width dimension, w=one wavelength at the frequency
Thickness, t=xc2xc wavelength at the frequency. 
Length dimension, L: Varies with type of array, noted below.
The element capacitance in Farads is given by Equation 1.
C=8.85xc3x9710xe2x88x9212*∈*w*L/txe2x80x83xe2x80x83Eq.1
Curve [82] is calculated for a typical 1 D array with element dimension L=32 wavelengths at the frequency. The signal loss varies from about 50% at 2.5 MHz to 85% at 15.0 MHz.
Curve [84] is calculated for a 1.25 D or 1.5 D array with element dimensions L=5.4 wavelengths at the frequency. The expected signal loss varies from about 82% at 2.5 MHz to 97% at 15.0 MHz. This loss may be intolerable, especially at the higher frequencies.
Curve [86] is calculated for a 1.75 D array with element dimensions L=2.5 wavelengths at the frequency. The expected signal loss varies from 91% at 2.5 MHz to 99% at 15.0 MHz. Curve [88] is calculated for a 2D array with element dimensions L=1.0 wavelengths at the frequency. The expected signal loss varies from 96% at 2.5 MHz to 99.5% at 15.0 MHz. These levels of signal loss in the cable are probably intolerable.
In the transmitting direction, the signal loss can be compensated by increasing the transmitter voltage. On reception, however, the loss in signal results in a decrease in signal to noise ratio. The resulting decrease in signal to noise ratio requires the use of lower frequencies or sacrificing the imaging depth into the body. It is possible to minimize this signal loss without using active electronics in the probe by inserting combinations of inductors and capacitors to impedance match the transducer element to the coaxial line impedance. This is burdensome when there are a hundred elements. With thousands of elements, this presents significant manufacturing problems as well as consuming space within the probe itself.
Referring back to FIG. 1 and to FIG. 4, there is shown in FIG. 4 a beamwidth illustration and comparison as between a linear array and a 1.5 D array. In the probe behind the flex circuit is an acoustical backing [23] that provides mechanical support and acoustical attenuation. When a piezoelectric transducer array [8] is electrically pulsed, two acoustical pulses are generated that travel in opposite directions. Pulse [27] traveling out of the scanhead into the target medium is desired, while the oppositely directed pulse propagating into the backing is unwanted and is absorbed by the backing.
One or more xe2x80x9cmatchingxe2x80x9d layers [26] are next in the path of pulse [27]. They serve to improve the coupling of energy from transducer array [8] into the body by matching the higher acoustical impedance array to the lower acoustical impedance of the target medium or body. This matching layer functions in the same way as the anti-reflection coating on an optical lens. The system electronics xe2x80x9cfocusxe2x80x9d the pulse in the scanning plane (or x-z plane) dimension.
Acoustic lens [6] is a simple convex lens as described in Erikson""s xe2x80x9cFocused Contact Transducer, U.S. Pat. No. 3,387,604, issued 1965 that forms the front surface that contacts the patient""s skin. It provides a fixed focus [33] to the sound pulse in the xe2x80x9cout-of planexe2x80x9d dimension, which is perpendicular to the scanning plane. Modern systems impose increasingly stringent requirements on arrays. As the number of transducers increases and their size decreases, however, the existing approaches may no longer be feasible or practical. Processing time, touch labor, yield, reliability and cost become limiting issues and new processes are required.
The first matrix array with an integrated circuit was developed by Erikon and Zuleeg in the early 1970""s. An 8xc3x978 element, 3.5 MHz receiver array with a preamplifier integrated circuit bump-bonded directly behind each 1 mmxc3x971 mm, Lithium Niobate, single crystal piezoelectric transducer was constructed. Individual transducers could be connected along a row to one of the eight output lines. Although the state of microelectronics was primitive by current standards, the array was shown to have acceptable sensitivity and demonstrated the feasibility of the approach. At that time, the diagnostic ultrasound industry was in its infancy and there was no need for such an array.
In 1995, Thomson Microsonics (TMX), now Thales Microsonics, Sophia-Antipolis, France, described a 1.25 D array, although the paper was entitled a 1.5 D array. This array did not have active circuitry in the scanhead. Individual transducers were connected through a separate coax cable.
More recently, TMX proposed a 128xc3x977 transducer curved linear array. Despite the title, this was also a 1.25 D array, by the definition adopted here. Active switching electronics in the scanhead were described, but no additional electronics were mentioned. A flex circuit was used for the interconnections between the array transducers and the electronics. There was no integrated circuit directly connected to the array. Signals are multiplexed through 128 coax lines.
In the U.S. government funded BUDI (Battlefield Ultrasonic Diagnostic Imager) program, a real-time, three dimensional ultrasound camera, intended for Army medics in combat situations, was designed and feasibility was proven. In this camera, an acoustical lens was used to image a volume onto a 128xc3x97128 (16,384 element) 5 MHz matrix array, as was disclosed in Erikson et al""s, xe2x80x9cImaging with a 2D Transducer Hybrid Arrayxe2x80x9d, in Acoustical Imaging, Vol. 24, Ed. H. Lee, Plenum Publishing Corp., New York, 2000. Each transducer of the piezocomposite array had a custom integrated circuit bump-bonded directly behind it using micro-solder balls. The piezocomposite array was air-backed, in other words there was a small air space between the array and the IC. The bump bonds were the only mechanical and electrical connections between the array and the IC. No matching layer was used on the front side of the array.
Each unit cell of the ROIC (read only integrated circuit) contained a preamplifier, signal processing, a limited amount of sampled data storage and multiplexing. The silicon was two side-buttable, permitting tiling of four, 64xc3x9764 pieces into a square 128xc3x97128 array.
The water tank images that were made demonstrated the viability of the acoustical lens and the performance of the receiver array. Most importantly, the concept of bump-bonding a matrix array directly to an integrated circuit was revalidated. Important parts of this technology are now protected by U.S. and foreign patents including Butler, N., et al, xe2x80x9cTwo Dimensional Transducer Integrated Circuitxe2x80x9d, U.S. Pat. No. 5,483,963, issued Jan. 16, 1996; White, T., et al, xe2x80x9cUltrasonic Array with Attenuating Electrical Interconnectsxe2x80x9d, U.S. Pat. No. 5,732,706, issued March 1998; and Erikson, K. R., et al, xe2x80x9cUltrasonic Cameraxe2x80x9d, U.S. Pat. No. 6,159,149, issued December 2000. This system has long term potential for diagnostic ultrasound applications; however, it is not useful for current medical ultrasound systems.
In summary, recent advances in transducer technology and integrated circuit fabrication have clearly extended the potential for improvements and extensions in the manner in which ultrasound applications are implemented.
One embodiment of the present invention provides an orthogonally switchable matrix array transducer system for ultrasound imaging. The system includes a multiplicity of subarrays of N by N transducers, and a bus including N vertical bus lines and N horizontal bus lines. N is at least 2. Each of the transducers of at least one subarray is switchably connected to a vertical bus line and to a horizontal bus line, thereby enabling reconfiguration of the at least one subarray between a horizontal mode of operation and a vertical mode of operation. A summer circuit may also be included which is adapted for integrating signals output by one or more of the subarrays into a common output signal that can be processed by a control system.
Another embodiment of the present invention provides a method for performing ultrasound imaging with a matrix array transducer probe. The transducer probe includes a multiplicity of tiled subarrays of transducers. The subarrays are oriented in a pattern having a vertical component overlapping a horizontal component, and at least one of the subarrays is switchable between a horizontal operating mode and a vertical operating mode. Output signals of the tiled subarrays can be integrated into an output signal. The method includes recording one or more ultrasound images in the horizontal operating mode, switching between the horizontal operating mode and the vertical operating mode, and recording one or more ultrasound images in the vertical operating mode. Note that the recording steps can be performed in any order. The method may further include integrating the ultrasound images recorded in the horizontal operating mode with the ultrasound images recorded in the vertical operating mode so as to generate real-time, three dimensional ultrasound imagery. In addition, the switching between the modes may be performed in real-time.
Another embodiment of the present invention provides an orthogonally switchable matrix array transducer system. The system includes a multiplicity of subarrays of transducers configured as orthogonally oriented arrays that include an area of overlap. The system further has a bus that includes vertical bus lines and horizontal bus lines, where each of the transducers of at least one subarray is switchably connected to a vertical bus line and to a horizontal bus line, thereby enabling real-time reconfiguration of the at least one subarray between a horizontal mode of operation and a vertical mode of operation. A summer circuit may also be included which is adapted for integrating signals output by one or more of the subarrays into a common output signal.
The features and advantages described herein are not all-inclusive and, in particular, many additional features and advantages will be apparent to one of ordinary skill in the art in view of the drawings, specification, and claims. Moreover, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes, and not to limit the scope of the inventive subject matter.